Grain-Boundary Topology and Motion; Deriving the Zener Limiting Grain Size; Minimizing Grain Growth in Nb-Microalloyed HSLA Steel: 21-Mat-A6 Materials Selection and Design for Materials Processing, December 2017, Q3

21-Mat-A6 Materials Selection and Design for Materials Processing · December 2017

Question 3: Grain-Boundary Topology and Motion; Deriving the Zener Limiting Grain Size; Minimizing Grain Growth in Nb-Microalloyed HSLA Steel (20 marks: a–2, b–6, c–8, d–4)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

3.1 — (a) Why six-or-more-sided grains grow, in a 2D polycrystalline network

In a fully annealed (strain-free) polycrystal, grain boundaries meet at triple junctions and mechanically equilibrate at 120 ° to balance the (equal, isotropic) boundary tensions pulling along each of the three boundaries. A grain with exactly SIX sides can, in principle, have every boundary perfectly straight while still meeting its neighbours at 120 ° at every corner (this is the geometry of a regular hexagon) — it is the unique polygon that satisfies both constraints (120 ° corners AND straight sides) simultaneously, so a 6-sided grain is in local topological equilibrium and its boundaries have zero net curvature (zero migration driving force). A grain with FEWER than six sides cannot hold 120 ° corners with straight sides — its boundaries must bow INWARD (concave, as seen from the grain's own interior) to make the angles work, and inward-curving boundaries migrate toward their own centre of curvature (into the small grain), so it shrinks. A grain with SIX OR MORE sides is forced the opposite way: its boundaries bow OUTWARD (convex from the interior) to satisfy the same 120 ° corner constraint over more sides, and outward-curving boundaries migrate away from the grain's own centre of curvature — i.e. into the neighbouring (smaller-sided) grains — so it GROWS at their expense. This is the topological (von Neumann–Mullins) statement of grain growth: at sufficiently high temperature (boundaries mobile), any configuration with six or more triple-junction intersections around a grain is geometrically compelled to have net-convex, migrating-outward boundaries, so growth is the inevitable, self-reinforcing consequence of the network's own triple-junction geometry, independent of any other driving force.

3.2 — (b) Two grain-boundary motion mechanisms with opposite migration sense

A curved grain boundary always has an intrinsic capillary (surface-tension) driving pressure toward its own centre of curvature, $P_\gamma=2\gamma_b/r$ for boundary energy $\gamma_b$ and local radius of curvature $r$ — this simply minimizes total grain-boundary area, exactly like surface tension pulling a soap film toward its centre of curvature. This capillary pressure is present in EVERY boundary, whether in a recrystallizing or a fully recrystallized (grain-growth) microstructure. The difference is what else acts on the boundary at the same time:

In grain growth, the driving force is capillarity ALONE — there is no stored deformation energy left (recrystallization is already complete, the matrix is strain-free on both sides of the boundary), so the only pressure available is $P_\gamma$, and the boundary simply migrates toward its own centre of curvature, shrinking small (few-sided) grains and growing large (many-sided) ones — exactly the mechanism in part (a).
In recrystallization, a NEW strain-free grain nucleus grows into a heavily strained, dislocation-dense matrix. The dominant driving force is now the difference in stored (dislocation) energy across the boundary, $\Delta G_v=E_v^{\text{deformed}}-E_v^{\text{strain-free}}\approx\tfrac12Gb^2\rho$, which is typically one to two orders of magnitude LARGER than the capillary pressure and always pushes the boundary AWAY from the low-energy (strain-free) side and INTO the high-energy (deformed) side — i.e. the boundary bulges outward, away from whatever local centre of curvature the capillary term alone would have dictated, because the stored-energy term overwhelms and can reverse the sign of the net local curvature response.

In short: the boundary always feels the same capillary pull toward its centre of curvature, but during recrystallization that pull is dominated and effectively overridden by a much larger stored-energy pressure pushing it the other way, into the more heavily deformed grain — the opposite migration sense the question describes.

3.3 — (c) Deriving the Zener limiting grain size $D_{\max}=4r/3f$

The grain-boundary migration driving pressure given in the question is $P_\gamma=2\gamma/D$. A dispersion of fine, incoherent precipitates of radius $r$ and volume fraction $f$ intersecting a unit area of boundary exerts an opposing, retarding pressure because each particle the boundary must cut through and re-form costs it boundary area, and hence energy, that would otherwise be eliminated by simply moving past. For a random dispersion, the number of particles intersected per unit boundary area is $N_A=3f/(2\pi r^2)$ (a standard stereological result for spheres of radius $r$, volume fraction $f$), and each intersected particle exerts a maximum pinning force $F_{\max}=\pi r\gamma$ (the boundary tension $\gamma$ acting around the particle's own great-circle perimeter $2\pi r$, resolved to its maximum retarding component). The resulting Zener pinning pressure is the force per unit area:

$$P_z = N_A\,F_{\max} = \frac{3f}{2\pi r^2}\times \pi r\gamma = \frac{3f\gamma}{2r}$$

Grain growth stalls (the final, limiting grain size $D_{\max}$ is reached) exactly when the driving pressure can no longer overcome the pinning pressure, i.e. when the two balance:

$$P_\gamma = P_z \quad\Rightarrow\quad \frac{2\gamma}{D_{\max}} = \frac{3f\gamma}{2r}$$

The interfacial energy $\gamma$ cancels (both pressures scale with the same boundary energy), leaving a purely geometric result:

$$\boxed{D_{\max} = \frac{4r}{3f}}$$

3.4 — (d) Minimum grain growth in Nb(C,N)-bearing HSLA steel

From $D_{\max}=4r/3f$, the final grain size is minimized by simultaneously making the precipitate radius $r$ as SMALL as possible and the volume fraction $f$ as LARGE as possible — i.e. maximizing the ratio $f/r$, a fine, densely dispersed particle population pins far more effectively than a coarse one of the same total volume. For a Nb(C,N)-microalloyed HSLA steel, this translates into three concrete processing conditions:

Keep the precipitates undissolved and stable at the reheat/rolling temperature. Nb(C,N) has a very low solubility product in austenite and does not coarsen (Ostwald-ripen) appreciably below about 1150 °C, unlike coarser, more soluble carbonitrides — so austenitizing/reheating BELOW the Nb(C,N) solvus temperature keeps $r$ small and $f$ at its maximum equilibrium value.
Maximize the as-precipitated volume fraction $f$ by alloying with the stoichiometric Nb content the steel's C+N content can support, and by precipitating at a temperature/time that favours fine, closely spaced particles (e.g. strain-induced precipitation during controlled rolling) rather than a smaller number of coarse ones.
Limit total time at high temperature once precipitation is complete, since even a stable dispersion will slowly coarsen (increasing $r$, and $D_{\max}$ with it) given long enough hold times — this is why controlled-rolling schedules for HSLA steels are time-and-temperature limited, not just temperature limited.

Together, these keep $D_{\max}=4r/3f$ small throughout reheating and hot rolling, which is precisely why Nb(C,N) is added to HSLA steels — to pin the austenite grain size fine before transformation, refining the final ferrite/pearlite grain size and improving both strength (Hall–Petch) and toughness.

Concept & Theory

Topic. Topological (von Neumann–Mullins) grain growth, the two driving-force regimes for grain-boundary motion, and Zener pinning by microalloy precipitates.

Key relations. Capillary pressure $P_\gamma=2\gamma/D$; stored-energy pressure $\Delta G_v\approx\tfrac12Gb^2\rho$; Zener pinning pressure $P_z=3f\gamma/2r$ (stereological particle intersection density $N_A=3f/2\pi r^2$ times per-particle pinning force $\pi r\gamma$); Zener limiting size $D_{\max}=4r/3f$.

Why this works. Every boundary-migration phenomenon in this question is a competition between the same handful of pressures (capillarity, stored strain energy, particle pinning); which one dominates, and hence which direction the boundary moves and how far growth proceeds, depends only on which driving forces are actually present and how finely/densely any pinning dispersion is distributed.

Common pitfall. Assuming a HIGHER precipitate volume fraction always helps regardless of particle size — a coarse dispersion at the same $f$ pins far less effectively ($D_{\max}\propto r$), so processing that lets Nb(C,N) coarsen (e.g. reheating above its solvus, then slow-cooling and re-precipitating coarse) can leave a LARGER limiting grain size than a finer dispersion at lower $f$.

Source. Callister & Rethwisch, Ch. 7.13–7.14 (grain growth, Zener pinning); Humphreys & Hatherly, Recrystallization and Related Annealing Phenomena, Ch. 9 (particle pinning, stereology); Gladman, The Physical Metallurgy of Microalloyed Steels, Ch. 2 (Nb(C,N) grain-boundary pinning in HSLA steel).